What Is AI’s Full Supply Chain Potential?
Brian Hoey - September 17, 2020
In a recent article about the factors that will impact the future of the supply chain, Gartner noted that while supply chain leaders really ought to have at least a 10 year plan for grappling with the transformation of the industry, most businesses don’t look beyond the next 3 years when making decisions. On the one hand, it’s not hard to see why a planner would be reluctant to make decisions with a truly long-term vision in mind—who really could have predicted the emergence of the coronavirus and its impact on the global value stream, for instance?
On the other hand, you do have to make some educated guesses about what the future will hold if you want to stay competitive. Why? Because the same old patterns only repeat themselves for so long, and eventually new trends always emerge. An automaker whose entire supply chain is organized around the idea that the automotive market will never change is going to have a difficult time as hybrid vehicles become more common and autonomous vehicles become feasible as consumer products, in the same way that manufacturers of all stripes who expected their suppliers to be in business forever are now finding themselves in a pickle.
Simply put, you need to have an idea of what the future of your supply chain will look like in order thrive in the present. From our perspective, that means that it’s time to start taking AI seriously.
The Current State of AI in the Supply Chain
Artificial intelligence and related technologies like machine learning, neural networks, etc. have already gone through too many hype cycles to count. They’re not yet fully mature technologies, which means that we’re still trying to envision exactly how they’ll slot into supply chain planning and other processes. As a result, a lot of commentary on this technology talks about the future of AI deployments, and how it might shape the evolving paradigms that will come to define the supply chains of the future. Like we saw above, this kind of forward-looking analysis can be important, but it also risks ignoring the concrete effects that new technologies are already having.
In the current supply chain, AI is already being adopted to power improved demand forecasting. Supply chain planners can use constraint programming to model complex vehicle routing problems, seeking out “good enough” solutions in areas where the optimal path would take too long to calculate. Inventory planners and shippers are using heuristics and metaheuristics to optimize 3D container loading, saving space and therefore money when storing and transporting product. By the same token, route and tour planners are using clustering to figure out the best way for deliveries to navigate multiple stops involving complex parameters for goods and timing. Though these applications don’t always seem futuristic, they’re crucial to creating leaner, more cost-efficient processes. In this way, they keep global businesses competitive even in the face of rapid change.
How AI Powers Industry 4.0
Though the implementations we just sketched out represent a real opportunity in themselves, they have the potential to be just the tip of the iceberg. Indeed, we expect AI to reach its full potential as a part of the emerging Industry 4.0 paradigms that will define the next evolution of the global supply chain. Stop us if you’ve heard this one before: Industry 4.0 will be built on unprecedented connectivity and transparency, leading to real-time workflows that involve autonomous machine decision-making and “cyber-physical systems” that introduce a new sophistication to the interactions between digitally-powered processes and physical network elements. Without AI, these paradigms would be impossible—AI is the only way to power lightning-fast analysis and decision-making processes that don’t always require the intervention of a human planner, just as it’s the only way to give human planners an understanding of systems that have too many data points to be grasped by back-of-the-envelope math or Excel macros.
In this way, we can begin to see how the full potential of AI in the supply chain involves ubiquitous, nearly invisible deployments of this kind of analytics technology. In the future, planners will take for granted that their APS software has built-in analytics to enable them to see the projected outcomes of any given decision. Likewise, logistics planners will come to expect self-optimizing transport networks that reroute trucks around traffic and weather incidents before they’ve even arisen. This will leave them with more time and attention to focus on the kinds of creative, personal, and strategic tasks that humans are best at—whether that’s dreaming up your next product launch, interfacing with suppliers to build relationships, or reimagining your supply and production networks to support further digital transformation.
From Here to There Eventually
Whether you’re only thinking on a 3 year timeline or you’re really trying to envision the supply chain of a decade or more into the future, there’s always going to be uncertainty. Uncertainty about what the future of AI in the supply chain will look like, yes, but also uncertainty about how navigate through transitional periods between technologies, paradigms, and changing market realities. With AI in particular, it’s crucial to ask yourself how you can adopt workflows and technologies that will both add value in the present and lay the groundwork for the future.
Pulling off this balancing act is going to partly technological and partly cultural. For starters, you’ll have to implement IT that lays a strong foundation for future AI deployments—which means working at every turn to increase connectivity and interoperability, whether that’s through a postmodern ERP approach or by adopting IT solutions explicitly designed to flexibly interwork with one another. You should have a strong roadmap with AI deployment milestones and KPIs for tracking your progress, and every new change should be in support of that roadmap. At the same time, you need to prime your personnel to think of AI as something that concretely adds value for their workflows, rather than something that’s going to make life harder. For this, explainable AI (i.e. predictive and prescriptive workflows that actually show users how decisions were reached, rather than just spitting out the “correct” answer) can be an important piece of the puzzle. By adopting solutions that are designed around helping the user make quick, well-informed decisions, you can get your people excited about an AI-powered future. Ultimately, that’s one of the most crucial steps to unlocking AI’s full potential in the supply chain.